Received: 9 September 2019 Accepted: 14 November 2019 DOI: 10.1111/jfb.14205

REGULAR PAPER FISH

Mode of miniaturisation influences body shape evolution in New World anchovies (Engraulidae)

Devin D. Bloom1,2 | Matthew Kolmann3 | Kimberly Foster1 | Helen Watrous1

1Department of Biological Sciences, Western Michigan University, Kalamazoo, Abstract Michigan, USA We explored the macroevolutionary dynamics of miniaturisation in New World 2 Institute of the Environment & Sustainability, anchovies by integrating a time-calibrated phylogeny, geometric morphometrics and Western Michigan University, Kalamazoo, Michigan, USA phylogenetic comparative methods. We found that the paedomorphic species 3Department of Biological Sciences, George Amazonsprattus scintilla occupies a novel region of shape space, while the dwarf spe- Washington University, Washington, DC, USA cies Anchoviella manamensis has an overall shape consistent with other anchovies. Correspondence We found that miniaturisation did not increase overall clade disparity in size or shape Devin D. Bloom, Department of Biological Sciences, Western Michigan University, 1903 beyond the expectations of Brownian motion, nor were there differences in rates of W. Michigan Avenue, Kalamazoo, MI size or shape evolution among clades. Overall, our study shows that while the mode 49008-5410, USA. Email: [email protected] of miniaturisation influences shape evolution, the phenotypic novelty produced by the evolution of miniaturisation did not seem to alter macroevolutionary dynamics. Funding information Directorate for Biological Sciences, Grant/ KEYWORDS Award Number: DEB 1754627 body size, Engraulidae, geometric morphometrics, macroevolution, miniaturisation, Neotropics

1 | INTRODUCTION with an increase in morphological variation and phenotypic novelty in a lineage (Hanken, 1986). Unlike developmentally truncated paedo- The evolution of extreme body size offers insight into the limits of morphic taxa, dwarf species resemble larger-bodied relatives (Britz phenotypic evolution in animals. The patterns and processes associ- et al., 2009; Ruber et al., 2007). However, the contribution of dwarf ated with large body size are well studied and evidence for selection species on the phenotypic diversity of a clade is largely unknown. If favouring large body size is omnipresent at both population (Roff, dwarf species retain the same bauplan and shape dimensions as larger 1991; Stearns, 1977) and macroevolutionary scales (Albert & Johnson, relatives, we predict they will reside in the same region of morphospace as non-miniaturised relatives. Under this scenario, dwarf species may not 2012; Jaffe et al., 2011; Pimiento et al., 2019; Slater et al., 2017; Ver- be associated with an increase in morphological variation. Alternatively, meij, 2012). Miniaturisation, or the extreme decrease in body size, has dwarf species may experience shifts that result in changes to the mor- evolved repeatedly in animals (Hanken & Wake, 1993) and a high phological diversity within a clade (Alberch et al., 1979; Britz & Conway, diversity of miniaturised fishes occur in the Neotropics (Goulding 2009; Britz et al., 2009; Buckup, 1993; Turner et al., 2007; Weitzman & et al., 1988; Roberts, 1984; Steele & López-Fernández, 2014; Vari, 1988) Few studies have explored the macroevolutionary patterns Weitzman & Vari, 1988). However, the processes shaping the evolu- and processes associated with different modes of miniaturisation nor tion of miniaturisation and the resulting phenotypic patterns are have directly compared the effect of modes of miniaturisation on pheno- poorly known (Blackenhorn, 2000), limiting our understanding of the typic evolution (Blackenhorn, 2000). role of extremely small body size in diversification. Miniaturisation has evolved numerous times in fishes, particularly in The evolution of miniaturisation can have a profound effect on South American freshwater rivers (Weitzman & Vari, 1988). In this phenotypic evolution and is often linked to morphological novelty study we use New World anchovies (Engraulidae) to explore the evolu- (Britz et al., 2009; Hanken, 1986; Yeh, 2002). There are two modes of tion of miniaturisation. New World anchovies are divided into two miniaturisation: paedomorphism and dwarfism. Paedomorphic species major clades: the marine clade and the South American clade (Bloom & typically arise due to shifts in the developmental process (Alberch Lovejoy, 2012). The marine clade includes approximately 45 species et al., 1979; Lovejoy, 2000) and are hypothesised to be associated and is dominated by near-shore, schooling and largely planktivorous

J Fish Biol. 2019;1–8. wileyonlinelibrary.com/journal/jfb © 2019 The Fisheries Society of the British Isles 1 2 FISH BLOOM ET AL. species found in both the eastern Pacific and western Atlantic Oceans. experiments or surgical procedures on live specimens. No part of this The speciose genus Anchoa Jordan & Evermann 1927 comprises most study involved live fishes. of the species diversity in the marine clade and along with several spe- cies of Anchoviella Fowler 1911, tend to have superficially similar exter- 2.1 | Morphometric data and analysis nal morphology and relatedly, seem to occupy similar ecological niches. This marine clade also includes several medium-sized, deep-bodied spe- We used body size data from Bloom et al. (2018). These authors com- cies in the genera Cetengraulis Günther 1868 and Anchovia Jordan & piled maximum LS body size data for all Clupeiformes from Fishbase Evermann 1895, as well as elongate species in the genus Engraulis (Froese & Pauly, 2017), Whitehead et al. (1988) and Carpenter (2002), Cuvier 1816. The South American clade, comprising six genera and which we pruned to focus on New World anchovies. Standard lengths more than 20 species, is ecologically diverse (Bloom & Egan, 2018; were log-transformed for subsequent analyses. For comparative ana- Bloom & Lovejoy, 2012; Egan et al., 2018a) and variable in size (Bloom lyses we used a published, multi-gene, time-calibrated phylogeny for et al., 2018). The majority of species in the South American clade are all Clupeiformes (Bloom & Lovejoy, 2014). The original phylogeny endemic to freshwater rivers in South America, but several species (e.g., included 52 New World anchovy species and was pruned to match Lycengraulis grossidens (Spix & Agassiz 1829) and Lycengraulis poeyi our size and shape data (described below; Figure 1). (Kner 1863)) are marine species that are derived from reversals back to Geometric morphmetrics is ideal for quantifying shape diversity of oceanic habitats and occur along the coasts of northern South America. miniaturised species because it characterises morphometric shape There are two species of miniaturised anchovies, Amazonsprattus independent of size, location and orientation (Angielczyk & Feldman, scintilla Roberts 1984 (<20 mm standard length; LS), a putatively paedo- 2013; Zelditch et al., 2004). Images of museum specimens of 45 spe- morphic species and Anchoviella manamensis Cervigón 1982 (<20 mm cies of Anchovies were taken from the left lateral side using a Nikon

LS), a dwarf species (Lavoué et al., 2010; Weitzman & Vari, 1988; D750 and a 60 mm macro lens (www.nikon.com). Bent, warped, or Whitehead et al., 1988). Roberts (1984) described A. scintilla,the juvenile specimens were not utilised to limit ontogenetic effects, smallest clupeiform, as a paedomorphic species based on the loss or shape distortion and other non-biological variation. Fourteen homolo- reduction of various characters, particularly the reduction of the pelvic gous, two-dimensional landmarks were used to summarise body shape girdle. Although the phylogenetic placement of A. scintilla was uncertain using the program TPSdig 2.31 (Rohlf, 2009). The landmarks (Figure 2) at the time of description, Lavoué et al. (2010) demonstrated that included: (1) anterior end of the rostral organ (snout); (2) dorsal A. scintilla was a member of New World anchovies and Bloom and orbit; (3) ventral orbit; (4) posterior of supraoccipital bone of the Lovejoy (2012) determined this species is a member of a clade of fresh- neurocranium; (5) dorsal tip of the operculum; (6) ventral tip of water anchovies from South America. Cervigón (1982) described the sub-operculum; (7) anterior insertion of the pectoral fins; (8) A. manamensis and hypothesised a phylogenetic affinity with other anterior insertion of pelvic fins; (9) anterior insertion of anal fin; freshwater and estuarine members of Anchoviella from north-eastern (10) posterior insertion of anal fin; (11) posterior insertion of dorsal South America. Several studies have clarified the phylogenetic relation- fin; (12) anterior insertion of dorsal fin; (13) dorsal insertion of caudal ships of New World anchovies and inferred the phylogenetic positions fin; (14) ventral insertion of caudal peduncle. We tested for a signifi- of both A. scintilla and A. manamensis (Bloom & Lovejoy, 2012, 2014; cant effect of size on shape using Procrustes regression with permuta- Lavoué et al., 2010, 2013). tion (nperm = 1000) (procD.allometry function in geomorph 3.0.5; In this study we investigated the patterns and processes of pheno- Adams et al., 2018). This method estimates the effect, or covariation, typic evolution associated with miniaturisation. We reconstructed of centroid size with our Procrustes-aligned shape coordinates, using body-size evolution across New World anchovies to determine a linear model (Collyer et al., 2015). whether miniaturisation evolves in primarily small bodied clades, or We used a phylomorphospace approach (Sidlauskas, 2008) to clades with high size disparity. We examined body shape diversity investigate patterns of body shape evolution among clades and modes across New World anchovies using geometric morphometrics to test of miniaturisation. If miniaturised taxa evolved convergent phenotypic whether the paedomorphic Amazonsprattus scintilla and dwarf Ancho- patterns then we expect that they would appear in proximate regions viella manamensis share similar phenotypes, or whether different of morphospace. We predicted that dwarf taxa would only be distin- modes of miniaturisation occupy different regions of body shape guishable from larger-bodied taxa according to their body size, but space. Finally, we used comparative methods to explore the macro- not in shape space, while paedomorphic taxa would be distinguishable evolutionary dynamics of New World anchovies and assess whether by both shape and size. We used the broken-stick method to inform the evolution of miniaturisation is associated with an increase in body how many principal component (PC) axes to retain for subsequent size and body shape disparity. analyses. In order to visualise a morphospace for anchovy body shape, we performed a phylogenetically-explicit PC analysis (using the phyl. 2 | MATERIALS AND METHODS pca function in phytools 0.6–99; Revell, 2012) to generate a scatterplot of PC scores and then projected the phylogeny onto this No fishes were collected as part of this study, as we used specimens morphospace using the plotGMPhyloMorphoSpace function in the from a series of natural history museums. We did not perform any geomorph package. BLOOM ET AL. FISH 3

Anchoa spinifer Jurengraulis juruensis Anchoviella manamensis Anchoviella guianensis 0652 Anchoviella carrikeri 3680 Anchoviella sp2 Anchoviella guianensis 3577 Anchoviella carrikeri 3694 Anchoviella alleni 0729 Anchoviella guianensis 0503 Anchoviella carrikeri 0806 Anchoviella brevirostris Anchovia surinamensis Lycengraulis batesii South American Clade Lycengraulis poeyi Lycengraulis grossidens Pterengraulis atherinoides Anchoviella lepidentostole Anchoviella jamesi Amazonspraus scinìlla Anchoviella alleni 0733 Anchoviella guianensis 3542 Anchoviella sp1 Anchoviella sp3 Anchoa filifera Cetengraulis edentulus Cetengraulis mysìcetus Anchoa nasus Anchoa lyolepis Anchoviella elongata Anchoa chamensis Anchovia macrolepidota Anchovia clupeoides Anchoa sp Anchoa colonensis Anchoa walkeri Anchoa schofieldi Anchoa panamensis Marine Clade Anchoviella balboae Anchoa mundeoloides Anchoa lamprotaenia Anchoa cubana Anchoa mitchilli Anchoa delicaìssima Anchoa parva Anchoa cayorum Engraulis mordax Engraulis japonicus Engraulis eurystole Engraulis encrasicolus Engraulis ringens Engraulis anchoita 10 cm 2.0 Body size (standard length) 26.0 Standard Length

FIGURE 1 Phylogeny of New World anchovies (Engraulidae) with body size (Standard length, LS cm) optimisd on the tree and the right panel shows LS ( ) of each species (modified from Bloom & Lovejoy, 2014)

of the dtt function in geiger (Pennell et al., 2014). These methods

12. 11. compare measured trait disparity (size or shape) relative to 10,000 4. 13. 5. Brownian motion (BM) simulations. The widely used morphological 2. 1. 10. 14. disparity index (MDI) suffers from issues with multiple testing and a 3. 7. 9. high false-positive rate, so we assessed deviations from the Brownian 6. 8. null using the rank-envelope test (Murrell, 2018). We were also interested in whether trends in body shape and body size traits evolved in such a way that they show either high or FIGURE 2 Landmarks used for geometric morphometrics of New World anchovies (Engraulidae) ( ): 1, anterior end of rostral organ; low phylogenetic signal. We calculated both Blomberg's K and Pagel's 2, dorsal orbit; 3, ventral orbit; 4, posterior of supraoccipital bone of λ using phytools (Revell, 2012). Values > 1.0 for K suggests that trait the neurocranium; 5, dorsal tip of the operculum; 6, ventral tip of variance is partitioned among clades, while values <1.0 suggest that sub-operculum; 7, anterior insertion of pectoral fins; 8, anterior insertion trait variance is found within clades. For λ, values nearest zero suggest of pelvic fins; 9, anterior insertion of anal fin; 10, posterior insertion of no correlation between species, while values nearer to one demon- anal fin; 11, posterior insertion of dorsal fin; 12, anterior insertion of dorsal fin; 13, dorsal insertion of caudal fin; and 14, ventral insertion strate a correlation between species that is equal to the Brownian of caudal peduncle expectation for signal.

2.2 | Phylogenetic comparative methods 3 | RESULTS To explore the tempo and mode of evolution of body size and body shape evolution in New World anchovies we employed a suite of Our reconstruction of body size confirms that different modes of comparative analyses. First, we reconstructed body size evolution miniaturisation evolved independently in anchovies. A. scintilla and across our phylogeny using a combined stochastic character mapping A. manamensis are members of separate clades of small-sized ancho- and maximum likelihood node estimation approaches, using the vies and both species originated early in the diversification of South simmap and ace functions in geiger (Pennell et al., 2014). We exam- American freshwater anchovies. Both miniaturised anchovy species ined how trends in body size and shape disparity changed through are closely related to other relatively small-bodied species; however, time by calculating subclade disparity and using disparity-through- the broader clade of South American anchovies includes a wide dis- time plots (DTT) with the function ddt1 (Murrell, 2018), a modification parity of body sizes (2–300 cm). 4 FISH BLOOM ET AL.

For body shape analyses, broken-stick method estimated that the use allometric scaling to correct our shape data, thereby retaining first five PC axes be retained for analysis, so these axes were used to any biological information contained therein. reconstruct DTT plots and phylomorphospaces. Negative loadings on Both the South American (P>0.05) and Marine clades (P>0.05) PC1 associated with elongate bodies, reduced jaw length and changes showed lower disparity in log size than the Brownian expectation in jaw position, as well as gracile caudal regions. Conversely, deeper- throughout most of their respective clade histories (Figure 4), but nei- bodied anchovies like Centengraulis Günther1868 and Anchovia ther result was significant. The South American clade does not show a clupeoides (Swainson 1839) loaded more positively on PC1 than other change in size around the stem age of miniaturised anchovies but taxa, with longer jaws and taller caudal peduncles. The dwarf anchovy, there is an increase as the clade approaches present day. Our analysis A. manamensis, has a body shape that places it in the same region of shape disparity through time showed that while the South Ameri- of morphospace as other members of New World anchovies, can clade (P>0.05) was consistent with a constant rates process, the including other species of Anchoviella and ecologically similar spe- Marine clade differed significantly from the Brownian expectation ciesinthemarinegenusAnchoa.WhileA. manamensis only (P < 0.01; Figure 5). Marine anchovies consistently had higher shape appeared distinct in body size space (Figure 3) and not in body disparity than the constant rates expectation, with peaks around shape space, Amazonsprattus was starkly delineated from other all 30 M year before present (MBP) and 10 MBP. Shape disparity in the other anchovies in both shape and size (Figure 3). The shape South American clade was not significantly different from the changes associated with the position of A. scintilla are a shorter jaw Brownian expectation, but peaked at approximately 20 MBP, a time and smaller mouth overall, as well as truncated anal fin relative to point that is roughly congruent with the stem ages of Anchoviella man- other anchovies. While we recovered a significant effect of cen- amensis (18 MBP) and Amazonsprattus scintilla (17 MBP). troid size on body shape for all New World anchovies (P <0.001), We found evidence for strong phylogenetic signal for body size we follow the rationale of others (Evans et al., 2019) and did not using both Pagel's method (λ = 0.98, P < 0.001and Blomberg's K

(a)

Anchoa panamensis Pterengraulis atherinoides Anchoa walkeri Amazonspraus scinìlla Anchoa mitchilli Lycengraulis batesii Lycengraulis grossidens Anchoa spinifer Anchoviella jamesi Anchoa cayorum Anchoviella carrikeri 3680 Anchoviella_lepidentostole Anchoa mundeoloides Anchoviella elongata Anchoviella manamensis - 18.7% variance Anchoviella brevirostris Anchoa parva

PC 2 Anchoa cubana Anchoa colonensis Anchovia clupeoides Anchoa filifera Anchoa lamprotaenia Anchovia surinamensis Jurengraulis juruensis Anchoa lyolepis

Anchoviella carrikeri 0806 Anchoviella sp3

Engraulis encrasicolus Engraulis mordax Cetengraulis mysìcetus Cetengraulis edentulus

Engraulis eurystole Engraulis ringens −0.05 0.00 0.05 0.10

−0.15 −0.10 −0.05 0.00 0.05 0.10 0.15

PC 1 - 28.4% variance

Pterengraulis atherinoides (b) Pterengraulis atherinoides Cetengraulis mysìcetus (c) (d) Anchoa lyolepis

0.05 Anchoa panamensis Amazonspraus scinìlla Engraulis eurystole Anchoa mitchilli Anchoa cubana Anchoa filifera Cetengraulis edentulus Engraulis mordax Engraulis ringens Anchoviella jamesi Lycengraulis batesii Anchoviella sp3 Anchoa_cayorum Anchovia clupeoides Anchoa lamprotaenia Anchoviella jamesi Engraulis ringens Anchoviella elongata Jurengraulis juruensis Anchoviella sp3 Anchoa walkeri Anchoa spinifer Anchoviella manamensis Anchoa parva Anchoa cubana Engraulis mordax Anchoa_mitchilli

- 12.3% variance Anchoa colonensis Lycengraulis grossidens Jurengraulis juruensis Anchoa colonensis - 9.9% variance Anchoviella carrikeri 3680 Cetengraulis mysìcetus Anchoviella carrikeri 3680 Anchoa_cayorum Anchoa_lamprotaenia - 6.7% variance Anchoviella_elongata Anchovia surinamensis Anchoa filifera Anchoa_mundeoloides Anchoa cubana Engraulis_mordax Anchoa walkeri Anchoa walkeri Anchoa_lyolepis Amazonspraus scinìlla Anchoviella brevirostris Anchoa parva Anchoa mundeoloides Lycengraulis_batesii Cetengraulis mysìcetus 0.00 PC 3 Anchoviella brevirostris Anchoviella manamensis Lycengraulis batesii

PC 4 Anchoviella jamesi Amazonspraus scinìlla Anchoa lamprotaenia Anchoa mundeoloides PC 5 Anchoa panamensis Anchoviella carrikeri 0806 Anchoa_parvaAnchoa_cayorum Anchovia clupeoides Engraulis eurystole Anchoviella lepidentostole Anchoa_mitchilli Pterengraulis atherinoides Cetengraulis edentulus Engraulis encrasicolus Anchovia surinamensis Engraulis encrasicolus Engraulis eurystole Anchoa panamensis Anchovia clupeoides Anchoviella elongata Anchoa spinifer Jurengraulis juruensis Anchoviella sp3 Anchoviella carrikeri 0806 Anchoa filifera Anchoviella lepidentostole Anchoviella carrikeri 0806 Anchoa colonensis Anchoa lyolepis Engraulis ringens Anchovia surinamensis Engraulis encrasicolus Anchoviella brevirostris Cetengraulis edentulus Anchoviella lepidentostole Anchoviella carrikeri 3680 Lycengraulis grossidens Lycengraulis grossidens −0.05 0.00 0.05 −0.05 0.00 0.05 −0.05

Anchoviella manamensis

Anchoa spinifer

−0.15 −0.10 −0.05 0.00 0.05 0.10 0.15 −0.15 −0.10 −0.05 0.00 0.05 0.10 0.15 −0.15 −0.10 −0.05 0.00 0.05 0.10 PC 1 PC 1 - 28.4% variance PC 1

FIGURE 3 Phylomorphospace for New World anchovies (Engraulidae) generated from principal component axes from geometric morphometrics: (a) phylomorphospace based on PC1 & PC2; (b) phylomorphospace based on PC1 & PC3; (c) phylomorphospace based on PC1 & PC4; (d) phylomorphospace based on PC1 & PC5. Members of the marine clade are black, members of the South America clade are grey, Anchoviella manamensis is blue, and Amazonsprattus scintilla is orange (modified from Whitehead et al., 1988) BLOOM ET AL. FISH 5

(c) All New World anchovies 1.5 (c) All New World anchovies 2.0 0.75 1.0 Disparity Disparity

(a) 2.0 (a) 0.0 0.0 0.0 1.0 0.0 0.5 1.0 3.0 0.5 Time (MBP) Time (MBP) 1.5

MDI: 0.002 2.0 P > 0.05 1.0 MDI: 0.06 P > 0.05 1.0 0.5 0.0

0.0 South American Clade (b) 1.2 (b) 1.2 Disparity Disparity 0.8 1.5 MDI: 0.009 P > 0.05 MDI: 0.21 P > 0.05 0.4 0.75

Marine Clade 0.0 0.0 40.0 20.0 0.0 Time (MBP) 40.0 20.0 0.0 Time (MBP) FIGURE 4 Disparity-through-time (DTT) plots of New World FIGURE 5 Disparity-through-time (DTT) plots of New World anchovies (Engraulidae) standard length over time for (a) the South anchovies (Engraulidae) body shape over time for (a) the South American clade ( ) and (b) the marine clade ( ). (c) Comparison American clade ( ) and (b) the marine clade ( ). (c) Comparison between these two clades and all New World anchovies. , between these two clades and all New World anchovies. , Brownian simulated expectation for mean clade disparity; MBP, Brownian simulated expectation for mean clade disparity; MBP, million years before present; MDI, morphological disparity index million years before present; MDI, morphological disparity index

(K = 0.90, P < 0.001). Tests of phylogenetic signal (Pagel's λ) for PC1 are comprised entirely of either dwarf or paedomorphic species, found that body shape is strongly predicted by the phylogeny (Frédérich et al., 2017; Lavoué et al., 2010; Takahashi & Ota, 2016), (λ = 0.99, P < 0.05), corroborated by similar results using Blomberg's K with the latter apparently the predominate mode of miniaturisation (K = 0.87, P < 0.05). Similarly, Pagel's λ results for PC2 indicate body (Hanken & Wake, 1993). Both A. scintilla and A. manamensis evolved shape is strongly predicted by the phylogeny (λ = 0.99, P < 0.001); early in South American anchovies and are embedded in clades of however, while Blomberg's method also shows a strong signal of the smaller-bodied anchovies (relative to marine taxa). Phylogenetic signal phylogeny, these results suggest that close relatives are even more for body size was strong, suggesting that miniaturisation did not nec- similar in size than expected based on Brownian (K = 1.15, P < 0.001). essarily involve major decreases in body size, but instead followed a continuous process of size reduction (Gould & MacFadden, 2004). 4 | DISCUSSION Clades generally have small-sized ancestors and increase in size over their evolutionary history (Albert & Johnson, 2012; Avaria-llautureo Our analysis of body size evolution demonstrates that two distinct et al., 2012; Guinot & Cavin, 2018). From our results for New World modes of miniaturisation, dwarfism and paedomorphism, evolved anchovies we infer a medium-sized ancestor that evolved both larger independently in South American anchovies. Miniaturisation has and smaller-bodied descendants, but the evolution of smaller extant evolved repeatedly throughout fishes (Weitzman & Vari, 1988) and species has been restricted to freshwater habitats. other fish clades (particularly in Otophysi) also exhibit repeated Extreme body size is generally predicted to evolve on islands, not instances of miniaturisation, including both dwarfism and paedomor- continents (Jaffe et al., 2011; Keogh et al., 2005; McClain et al., 2013; phism (de Santana & Crampton, 2011; Lavoué et al., 2008; Ruber Thomas et al., 2009; Wollenberg et al., 2011). Both miniaturised spe- et al., 2007). However, most lineages of miniaturised species cies in our analysis are freshwater (continental) residents and 6 FISH BLOOM ET AL. miniaturisation evolved more commonly in freshwater environments ramifications of miniaturisation. Indeed, there are cases where extreme in Clupeiformes (Lavoué et al., 2008; 2010; Roberts, 1972; Stiassny, body size has influenced macroevolutionary processes (Frédérich et al., 2002; Whitehead & Teugels, 1985). It is possible that certain habitats, 2017; Jaffe et al., 2011; Slater et al., 2017). Understanding the evolution such as riverine systems, function as island-like ecological settings of miniaturised Neotropical fishes will require fine-scale data on habitat (Mcclain et al., 2006) and offer similar ecological opportunity for occupancy, trophic niche and life history of these intriguing fishes. extreme body size evolution. Alternatively, Thomas et al. (2009) argued that extreme body size is not driven by islands per se, but ACKNOWLEDGEMENTS rather that when a lineage colonises an island, it is the novel environ- ment in general that promotes phenotypic evolution. Anchovies We thank the following institutions and curatorial staff for specimen colonised freshwater rivers of South America 20–30 MBP (Bloom & loans: H. Lopez-Fernandez, E. Holm, M. Burridge and D. Stacey (Royal Egan, 2018; Bloom & Lovejoy, 2017; Egan et al., 2018a), which pre- Ontario Museum), Chris Taylor (Illinois Natural History Survey), ceded or was synchronous with the Pebas mega-wetland (Hoorn J. Sparks (American Museum of Natural History), H. Lopez-Fernandez et al., 2010; Wesslingh & Hoorn, 2010). The Pebas mega-wetland was (University of Michigan Museum of Zoology), K. Hartel and a dynamic ecosystem that may have created a novel environment that A. Williston (Museum of Comparative Zoology, Harvard University). set the stage for diversification of some Neotropical freshwater fishes, Michael Burns provided R code and help with analyses. Nathan including freshwater anchovies (Bloom & Lovejoy, 2017). Island-like Lovejoy provided specimens and constructive feedback during the or not, new environments can drive shifts in body size evolution due early stages of this project. Jane Thayer and Michael Dobrovetsky to changes in competition, prey and other resource availability and photographed and digitised specimens used in this study. associated selection on life history traits (Palkovacs, 2003; Steele & López-Fernández, 2014; Zuanon et al., 2006). Extreme body size is often associated with phenotypic novelty, AUTHOR CONTRIBUTIONS particularly in paedomorphic species (Britz et al., 2009; Hanken, 1986; D.D.B. and M.K. conceived and designed the study. D.D.B., M.K., Hanken & Wake, 1993; Yeh, 2002). These novel phenotypes often K.F. and H.W. collected the data. M.K. and K.F. analysed the data. include the loss or pervasive reduction of traits; a remarkable example D.D.B. wrote the first draft of the manuscript and all authors contrib- is the diminutive cyprinid Danionella dracula Britz, Conway & Rüber uted to writing and preparation of the final draft. 2009 that lacks over 40 bones found in close relatives (Britz et al., 2009). Our results showed that the paedomorphic A. scintilla occupies novel body shape space for New World anchovies, supporting a gen- ORCID eral pattern of phenotypic novelty associated with this mode of Devin D. Bloom https://orcid.org/0000-0002-5799-5796 miniaturisation. We also found that the dwarf A. manamensis has an overall body shape that is consistent with other non-miniaturised anchovies and is not associated with an increase in phenotypic dispar- REFERENCES ity. Frédérich et al. (2017) showed that independent clades of marine Adams, D.C., Collyer, M.L. and Kaliontzopoupou, A. (2018). Geomorph: dwarf angelfishes (Pomacanthidae) experienced convergent evolution software for geometric morphometric analyses. R package ver- and diversification into new adaptive zones. However, our results sug- sion 3.0.5. gest that convergent body shape evolution is unlikely when lineages Alberch, P., Gould, S. J., Oster, F. G., & Wake, D. B. (1979). Size and shape in ontogeny and phylogeny. Paleobiology, 5, 296–317. independently evolve extreme body size reduction via different modes Albert, J., & Johnson, D. (2012). Diversity and evolution of body size in of miniaturisation. The convergent evolution of dwarf angelfishes fishes. Evolutionary Biology, 39, 324–340. seems to be associated with a common shift in ecology in each respec- Angielczyk, K. D., & Feldman, C. R. (2013). Are diminutive turtles mini- tive instance, a process that may not occur when different modes of aturised? The ontogeny of plastron shape in emydine turtles. Biological – miniaturisation arise. Indeed, A. scintilla and A. manamensis are rarely Journal of the Linnean Society, 108, 727 755. Avaria-llautureo, J., Canales-aguirre, C. B., Boric-bargetto, D., Morales- collected in the same microhabitat (D.D.B., pers. obs.) and these species pallero, B., & Rodrıguez, E. (2012). Body size evolution in extant probably occupy different niche spaces. While the functional advan- oryzomyini rodents: Cope's rule or miniaturisation? PLoS ONE, 7,1–8. tages of miniaturisation in Neotropical fishes remain unclear (Goulding Blackenhorn, W. U. (2000). The evolution of body size: What keeps organ- et al., 1988), predator avoidance and competitive exclusion are possible isms small? The Quarterly Review of Biology, 75, 385–407. Bloom, D. D., Burns, M. D., & Schriever, T. A. (2018). Evolution of body explanations. size and trophic position in migratory fishes: A phylogenetic compara- The presence of morphological novelty and extreme body size are tive analysis of Clupeiformes (anchovies, herring, shad and allies). Bio- factors that would seemingly prime a clade for high phenotypic dispar- logical Journal of the Linnean Society, 125, 302–314. ity. However, we found no evidence that miniaturisation influences Bloom, D. D., & Egan, J. P. (2018). Systematics of Clupeiformes and testing morphological disparity. Moreover, the evolution of extreme size reduc- for ecological limits on species richness in a trans-marine/freshwater clade. Neotropical Ichthyology, 16,1–14. tion did not alter macroevolutionary processes and patterns. However, Bloom, D. D., & Lovejoy, N. R. (2012). Molecular phylogenetics reveals a our study system includes only two origins of miniaturisation, which pattern of biome conservatism in New World anchovies (family limits our ability to draw broad inferences on the macroevolutionary Engraulidae). Journal of Evolutionary Biology, 25, 701–715. BLOOM ET AL. FISH 7

Bloom, D. D., & Lovejoy, N. R. (2014). The evolutionary origins of dia- Keogh, J. S., Scott, I. a. W., & Hayes, C. (2005). Rapid and repeated origin dromy inferred from a time-calibrated phylogeny for Clupeiformes of insular gigantism and dwarfism in Australian tiger snakes. Evolution, (herring and allies). Proceedings of the Royal Society B: Biological Sci- 59, 226–233. ences, 281, 20132081. Lavoué,S.,Miya,M.,Kawaguchi,A.,Yoshino,T.,&Nishida,M.(2008). Bloom, D. D., & Lovejoy, N. R. (2017). On the origins of marine-derived The phylogenetic position of an undescribed paedomorphic freshwater fishes in South America. Journal of Biogeography, 44, clupeiform taxon: Mitogenomic evidence. Ichthyological Research, 1927–1938. 55,328–334. Britz, R., & Conway, K. W. (2009). Osteology of Paedocypris, a miniature Lavoué, S., Miya, M., Musikasinthorn, P., Chen, W.-J., & Nishida, M. and highly developmentally truncated fish (Teleostei: Ostariophysi: (2013). Mitogenomic evidence for an Indo-West Pacific origin of the Cyprinidae). Journal of Morphology, 270, 389–412. Clupeoidei (Teleostei: Clupeiformes). PLoS ONE, 8, e56485. Britz, R., Conway, K. W., & Rüber, L. (2009). Spectacular morphological Lavoué, S., Miya, M., & Nishida, M. (2010). Mitochondrial phylogenomics novelty in a miniature cyprinid fish, Danionella dracula n. sp. Proceed- of anchovies (family Engraulidae) and recurrent origins of pronounced ings of the Royal Society B, 276, 2179–2186. miniaturisation in the order Clupeiformes. Molecular Phylogenetics and Buckup, P. A. (1993). Phylogenetic interrelationships and reductive evolu- Evolution, 56, 480–485. tion in Neotropical characidiin fishes (Characiformes, Ostariophysi). Lovejoy, N. R. (2000). Reinterpreting recapitulation: Systematics of Cladistics, 9, 305–341. needlefishes and their allies (Teleostei: Beloniformes). Evolution, 54, Carpenter, K. E. (2002). The living marine resources of the Western Central 1349–1362. Atlantic. Vol. 2. Bony Fishes part 1 (Acipensideridae to Grammatidae). McClain, C. R., Boyer, A. G., & Rosenberg, G. (2006). The island rule and Rome, FAO FAO Species Identification Guide for Fishery Purposes the evolution of body size in the deep sea. Journal of Biogeography, 33, and American Society of Ichthyologists and Herpetologists Special 1578–1584. Publication No. 5. McClain, C. R., Durst, P. a. P., Boyer, A. G., & Francis, C. D. (2013). Cervigón. (1982). Los recursos pesqueros del rio Orinoco y su exploitacion Unravelling the determinants of insular body size shifts. Biology Letters, (p. 386). Caracas, Venezuela: Corporacion Venezolana de Guayana, 9, 20120989. Division de Desarrolo Agricola. Murrell, D. J. (2018). A global envelope test to detect non-random bursts Collyer, M. L., Sekora, D. J., & Adams, D. C. (2015). A method for analysis of trait evolution. Methods in Ecology and Evolution, 9, 1739–1748. of phenotypic change for phenotypes described by high-dimensional Palkovacs, E. P. (2003). Explaining adaptive body size on islands: A life his- data. Heredity, 115, 357. tory appoach. Oikos, 103,37–44. de Santana, C. D. D. S., & Crampton, W. G. R. (2011). Phylogenetic interre- Pennell, M. W., Eastman, J. M., Slater, G. J., Brown, J. W., Uyeda, J. C., lationships, taxonomy and reductive evolution in the Neotropical elec- Fitzjohn, R. G., … Harmon, L. J. (2014). Geiger v2.0: An expanded suite tric fish genus Hypopygus (Teleostei, Ostariophysi, Gymnotiformes). of methods for fitting macroevolutionary models to phylogenetic Zoological Journal of the Linnean Society, 163, 1096–1156. trees. Bioinformatics, 30, 2216–2218. Egan, J. P., Bloom, D. D., Kuo, C.-H., Hammer, M. P., Tongnunui, P., Pimiento, C., Cantalapiedra, J. L., Shimada, K., Field, D. J., & Smaers, J. B. Iglésias, S. P., … Simons, A. M. (2018a). Phylogenetic analysis of trophic (2019). Evolutionary pathways toward gigantism in sharks and rays. niche evolution reveals a latitudinal herbivory gradient in Clupeoidei Evolution, 73, 588–599. (herrings, anchovies and allies). Molecular Phylogenetics and Evolution, Revell, L. J. (2012). phytools: An R package for phylogenetic comparative 124, 151–161. biology (and other things). Methods in Ecology and Evolution, 3, Evans, K. M., Williams, K. L., & Westneat, M. W. (2019). Do coral reefs 217–223. promote morphological diversification? Exploration of habitat effects Roberts, T. R. (1972). Osteology and description of Thrattidion noctivagus, on labrid pharyngeal jaw evolution in the era of big data. Integrative a minute, new freshwater clupeid fish from Cameroon, with a discus- and Comparative Biology, 59, 696–704. sion of Pellonulin relationships. Museum of Comparative Zoology, Frédérich, B., Santini, F., Konow, N., Schnitzler, J., Lecchini, D., & 382,2–25. Alfaro, M. E. (2017). Body shape convergence driven by small size Roberts, T. R. (1984). Amazonsprattus scintilla, a new genus and species optimum in marine angelfishes. Biology Letters, 13, 20170154. from the Rio Nego, Brazil, the smallest known clupeomorph fish. Pro- Froese, R., & Pauly, D. (2017). FishBase. Retrieved from www.fishbase.org ceedings of the California Academy of Sciences, 43, 317–321. Gould, G. C., & MacFadden, B. J. (2004). Gigantism, dwarfism and Cope's Roff, D. A. (1991). Life history consequences of bioenergetic and biome- rule: “Nothing in evolution makes sense without a phylogeny”. Bulletin chanical constraints on migration. American Zoologist, 31, 205–215. of the American Museum of Natural History, 285, 219–237. Rohlf, F.J., 2009. tpsDig2, Version 2.1. State University of New York, Goulding, M., Carvalho, M. L., & Gerreira, E. G. (1988). Rio Negro, rich life in Stony Brook. poor water. The Hague, Netherlands: SPB ACademic Publishing. Ruber, L., Kottelat, M., Tan, H., Ng, P., & Britz, R. (2007). Evolution of Guinot, G., & Cavin, L. (2018). Body size evolution and habitat colonization miniaturisation and the phylogenetic position of Paedocypris, comprising across 100 million years (Late Jurassic–Paleocene) of the actino- the world's smallest vertebrate. BMC Evolutionary Biology, 7,38. pterygian evolutionary history. Fish and Fisheries, 19, 577–597. Sidlauskas, B. (2008). Continuous and arrested morphological diversifica- Hanken, J. (1986). Morphological novelty in the limb skeleton accompanies tion in sister clades of characiform fishes: A phylomorphospace miniaturisation in salamanders. Science, 229, 871–874. approach. Evolution, 62, 3135–3156. Hanken, J., & Wake, D. B. (1993). Miniaturisation of body size: Organismal Slater, G. J., Goldbogen, J. A., & Pyenson, N. D. (2017). Independent evolu- consequences and evolutionary significance. Annual Review of Ecology tion of baleen whale gigantism linked to Plio-Pleistocene ocean and Systematics, 24, 501–519. dynamics. Proceedings of the Royal Society B, 284, 20170546. Hoorn, C., Wesselingh, F. P., ter Steege, H., Bermudez, M. A., Mora, A., Stearns, S. C. (1977). The evolution of life history traits: A critique of the Sevink, J., … Antonelli, A. (2010). Amazonia through time: Andean theory and a review of the data. Annual Review of Ecology and System- uplift, climate change, landscape evolution and biodiversity. Science, atics, 8, 145–171. 330, 927–931. Steele, S. E., & López-Fernández, H. (2014). Body size diversity and fre- Jaffe, A. L., Slater, G. J., & Alfaro, M. E. (2011). The evolution of island quency distributions of Neotropical cichlid fishes (Perciformes: gigantism and body size variation in tortoises and turtles. Biology Let- Cichlidae: Cichlinae). PLoS ONE, 9, e106336. https://doi.org/10.1371/ ters, 7, 558–561. journal.pone.0106336. 8 FISH BLOOM ET AL.

Stiassny, M. L. J. (2002). Revision of Sauvagella Bertin (Clupeidae; Whitehead, P. J. P., Nelson, G. J., & Wongratana, T. (1988). FAO species Pellonulinae; Ehiravini) with a description of a new species from the catalogue Clupeoid fishes of the world (Suborder Clupeoidei) (Vol. 7). freshwaters of Madagascar and diagnosis of the Ehiravini. Copeia, Rome, Italy: UNDP FAO. 2002,67–76. Wollenberg, K. C., Vieites, D. R., Glaw, F., & Vences, M. (2011). Speciation Takahashi, T., & Ota, K. (2016). Body size evolution of a shell-brooding in little: The role of range and body size in the diversification of Mala- cichlid fish from Lake Tanganyika. Journal of Evolutionary Biology, 29, gasy mantellid frogs. BMC Evolutionary Biology, 11, 217. 2373–2382. Yeh, J. (2002). The effect of miniaturised body size on skeletal morphology Thomas, G. H., Meiri, S., & Phillimore, A. B. (2009). Body size diversifica- in frogs. Evolution, 56, 628–641. tion in Anolis: Novel environment and island effects. Evolution, 63, Zelditch, M. L., Swiderski, D. L., Sheets, H. D., & Fink, W. L. (2004). Geo- 2017–2030. metric morphometrics for biologists. London, UK: Elsevier Academic Turner, A. H., Pol, D., Clarke, J. A., Erickson, G. M., & Norell, M. A. (2007). Press. A basal dromaeosaurid and size evolution preceding avian flight. Sci- Zuanon, J., Bockmann, F. A., & Sazima, I. (2006). A remarkable sand- ence, 317, 1378–1381. dwelling fish assemblage from central Amazonia, with comments on Vermeij, G. J. (2012). The evolution of gigantism on temperate seashores. the evolution of psammophily in South American freshwater fishes. Biological Journal of the Linnean Society, 106, 776–793. Neotropical Ichthyology, 4, 107–118. Weitzman, S. H., & Vari, R. P. (1988). Miniaturisation in South American freshwater fishes; an overview and discussion. Proceedings of the Bio- logical Society of Washington, 101, 444–465. Wesslingh, F., & Hoorn, C. (2010). Amazonia, landscape and species How to cite this article: Bloom DD, Kolmann M, Foster K, evolution: A look into the past. West Sussex, UK: Blackwell Watrous H. Mode of miniaturisation influences body shape Publishing. evolution in New World anchovies (Engraulidae). J Fish Biol. Whitehead, P. J., & Teugels, G. G. (1985). The West African pygmy herring – Sierrathrissa leonensis: General features, visceral anatomy and osteol- 2019;1 8. https://doi.org/10.1111/jfb.14205 ogy. American Museum Novitates, 2835, 1481.